Biomedical Engineering Reference
In-Depth Information
Figure 9. Regions displaying significant longitudinal gray matter atrophy over a 4-year
period. Estimates of longitudinal atrophy were determined by segmentation into GM,
WM and CSF, then applying the mass-preserving RAVENS methodology described in the
text, which deforms each individual's brain into alignment with a template brain, while
preserving tissue mass by converting it to density. Voxel-based analysis of the resultant
tissue density maps is equivalent to voxel-based volumetric analysis and therefore of atro-
phy quantification. Reprinted with permission from [3]. Copyright c
2001, Society for
Neuroscience. See attached CD for color version.
above has been applied to structural images from this study, in order to measure
patterns of significant longitudinal atrophy in normal healthy adults. Figure 9
displays a 3D rendering of the regions displaying significant longitudinal atrophy
over a 4-year period.
In addition to atrophy and other shape changes, signal changes are very pro-
nounced even in normal aging. Most importantly, the white matter tends to get
darker in standard T1-SPGR images, perhaps due to underlying vascular disease,
demyelination, mineral deposition, or other degenerative processes. The meth-
ods described in this chapter can also be used in voxel-based analysis of signal
changes. In [1] we studied WM darkening in the BLSA sample. Figure 10 dis-
plays a statistical parametric map of voxel-wise analysis of longitudinal change.
Quantification of signal changes is just as important as measurement of atrophy,
for many reasons. First, many degenerative changes can be first manifested as
signal changes, prior to tissue atrophy. Second, in early AD stages, vascular dis-
ease or demyelination of WM can have an additive effect on cognitive decline
to that of AD pathology. Therefore, structural and functional changes caused by
vascular disease, and which most often change MR signal characteristics, must be
well characterized in order to isolate the imaging signatures that are specific to
AD. Moreover, our previous work suggests that, from certain perspectives, signal
changes carry orthogonal information to atrophy [1].
3.6.2. Neuroimaging of schizophrenia
We have performed voxel-based analysis of the RAVENS tissue density maps
in a study of 158 normal controls and schizophrenia patients. Using voxel-based
analysis of the RAVENS maps, we have identified spatially complex patterns of
tissue volume differences between healthy controls and patients. Moreover, using
Search WWH ::




Custom Search